Spectacular Bantu tone

Michael Marlo and Kris Ebarb, University of Missouri at Columbia)

In this talk we present vignettes of some of the most striking and theoretically significant patterns we have encountered in our research on the tonal systems of Bantu languages. Tone distinguishes lexical items, and is a major feature of the marking of tense-aspect-mood-polarity (TAMP) and clause type distinctions in Bantu languages. We will discuss data from Kuria (Kenya, Tanzania), Bakweri (Cameroon), and the Khayo and Idakho varieties of Luyia (Kenya) showing the diverse realization of tonal patterns that mark TAMP distinctions. These so-called ‘melodies’ can generally be characterized as the addition of one or two tones, usually but not always a H tone, to one of the edges of the verb stem.

In Kuria, H tones are assigned to one of the first four moras of the verb stem, depending on TAMP properties of the verb. Phonological systems that count to four are quite rare and potentially theoretically problematic. A second surprising feature of Kuria tone is that although tone assignment is calculated from the left edge of the verb stem, a unit of structure within the verb, tone can be assigned to the word that follows the verb. This challenges theories that claim that word-internal morphological structure is inaccessible at the level of the phrase.

Bakweri is quite different from Kuria. All inflectional tones are assigned to a single position within the verb, the final vowel, but are distinguished by the tone or combination of tones that is assigned: H, L, Ø, HL, or LH. Two highlights of this section are (i) that tonal representations (e.g. H vs. L vs. Ø) matter, and (ii) selecting the correct tonal melody to place on the verb requires information not only about TAMP properties but also other morphosyntactic information such as whether the verb is in a relative clause.

In the Khayo and Idakho varieties of Luyia, accounting for differences in the tonal patterns of different verb tenses requires construction-specific rules which apply in some tenses but not others. Not only does the location of melodic tones distinguish verb tenses, as in Kuria, but also whether additional tonal rules, such as spreading or deletion rules, apply in those tenses.

An additional striking pattern of Idakho verbal tonology concerns the complex realization of a H tone associated with the passive suffix. This tone has an unusual dependency relationship with inflectional tones and aspects of the verb morphology which raises questions for the nature of the phonology-morphology interface.